Browsing by Author "BRAHMI, Raid"
Now showing 1 - 1 of 1
- Results Per Page
- Sort Options
Item Embedded System for Corrosion Analysis and Prediction(University of Echahid Cheikh Larbi Tébessi -Tébessa, 2025-06-09) BRAHMI, RaidCorrosion is a major challenge in the oil and gas industry, compromising the safety, reliability, and sustainability of transportation infrastructure. This project proposes the design of an embedded system dedicated to analyzing and predicting corrosion in metal pipelines. It is based on the exploitation of data from inspections carried out over several years, combined with a BiLSTM (Bidirectional Long Short-Term Memory) deep learning model to predict the maximum corrosion depth at a given horizon. The dataset undergoes various preparation steps: missing value processing, correlation analysis and feature matching between inspection years. The developed model demonstrated high predictive accuracy with a low mean absolute error. This approach enables effective predictive maintenance and contributes to failure prevention, thus providing an innovative and suitable solution for real-time corrosion monitoring in embedded industrial systems.